Intel to spend $20B on two new chip factories in Arizona
nytimes.com
nytimes.com
https://en.wikipedia.org/wiki/List_of_semiconductor_fabricat...
The US would fight to prevent China taking over Taiwan by force, and Japan and others would assist them, but if China succeeded, the US wouldn't choose to start the third world war by using weapons carelessly... it would accomplish nothing at all.
I found that point really interesting - and also perplexing.
I can understand why US companies would choose to outsource manufacturing concerns to East Asian countries, but I'm wondering why the opposite is true in this case here.
Could someone with insight into these matters please help me understand - why would TSMC or Samsung choose to build factories in the US?
Government subsidies on the upfront CapEx aside - I'm curious what the long term benefits to them would be.
Thanks :)
If China has hostile intent it also increases the risk of Chinese action, because - dependent on stockpiles and security - Taiwan is a resource choke point which could potentially decapitate the entire US military machine.
This is a terrible situation for the US to be in.
As the US replaces that capability on its home turf Taiwan's strategic value wanes, which paradoxically makes an attack less likely because China can get what it wants from Taiwan by conventional economic and political means.
The question is... how lucky/aggressive does China feel? And how just-in-time is the US military supply chain?
The military doesn't really worry much about price.
It's not just a price component, but a performance and viability of new designs one, at a time where the DoD internally is pushing ML hard as a major pillar of investment in weapons technology.
If you're TSMC or Samsung you probably had a conversation with US officials along the lines of: Build fabs in the US, or we will tariff and tax our way back to a mostly domestic, American owned industry. 60% of TSMC's sales are to the US customers. There's not much of a choice there.
Instead, this move ties the success of the American facility and the Taiwanese facility together, building a bond and interdependence.
Sort of.
TSMC's interests do not perfectly align with Taiwan's. TSMC putting fabs in the well-defended territory of its customers makes sense.
That said, China being able to credibly threaten a critical supply line is a security issue for America. It constrains U.S. response options to an invasion, threatened or acted upon. Removing that constraint does make Taiwan a bit less critical to America. But U.S.-Taiwanese relations predate TSMC. On the net, these plants free America to be more aggressive with China, which is broadly in Taiwan's interest.
Why now? In a word: automation. Semiconductor manufacturing has become far less labor intensive since TSMC was founded in 1987. That makes the American labor premium less material.
I think you're failing to take into account a number of things. First and foremost the very foundation of the US is based on respecting IP law given we're almost entirely an "information" society. If we suddenly steal TSMC's business secrets because it's convenient, what's preventing Europe and our allies from taking ours? Why would they respect IP law if we don't?
Second, regardless of the technical reliance we have on Taiwan, there's no planet on which the US wants China expanding and taking over other countries in Asia for political reasons.
The factories in the US ensure that the US keeps sending business to TSMC.
It probably comes with some benefits, but I doubt volunteering to be a potential hotspot for WW3 is really worth those benefits.
But one thing is certain - it's a question of when, not if. Taking control of Taiwan is too important to the founding myth of the CCP. Every reminder that Taiwan exists is a challenge to their power. Their desire for Taiwan has never wavered in the last 72 years, nor do I think it will ever. Meanwhile Taiwan's main defender (America) has no permanent allies, only permanent interests. Once the semiconductor supply chain is replicated successfully in the US, they might find that they're no longer willing to sacrifice American lives to protect the sovereignty of Taiwan. At some point in the next 10-40 years, Taiwan will be a bargaining chip America will trade away for something else they value more.
Not saying that you're wrong, because you seem much more informed and involved on/in this subject than I am, but from half the world away I can say that around 10 years or so ago that "certainty" wasn't that "certain", so to speak, at least not from what I could gather by reading the "local" press (mostly the Financial Times and The Economist).
Do you happen to have any resources available that further comment on that certainty? Again, not because I'm thinking that you're wrong, but because I'm really curious about this subject (there was a similar Taiwan-related post on the front-page of this website a few days ago but it dropped off pretty quickly).
> To many Chinese, the island’s conquest is a sacred national mission
Taiwan is definitively on the next because it is way too close in distance to China for its own good.
No, I was looking for sources that would talk about the "certainty" of invasion by China (in a large enough interval, from let's say tomorrow to 10-20 years into the future), like I said, about 10 years ago that was not the official discourse.
"Predicting the future" would be to ask for a specific date/time interval for that supposed invasion, which I did not ask for because that would have been a reasonably difficult task to predict, I concur.
The US has a permanent interest in the continued existence of an independent, non CCP aligned Taiwan. The only thing that time and time again frustrated the ambitions of Great Powers has been geography. Taiwan is the cork in the bottle containing the PLAN within the first island chain, without which China can operate freely throughout the Pacific. Just about every projection has China become the world's largest economy. America's only play to maintain dominance is to make sure that the tyranny of geography restrains the Chinese from operating militarily as freely as they would want. Much American blood had been shed in winning the Pacific in the first place. America is willing to shed more blood to keep it that way.
imposing control on Hong Kong happened because it did not require high-stakes combat with a nuclear peer competitor
I'd bet on it happening during Xi's reign, as it's critical to his pursuit of legacy. He's obsessed with writing his own name into China's history, taking Taiwan (and simultaneously defeating the West as he does it) would be a very big stamp on that. It would formally signal China's new empire-like dominion over Asia.
I don't think the US will trade Taiwan as a bargaining chip. By the time the US is willing to trade it, China will fully understand it's not really worth bargaining over any longer. The time to bargain would be now, or in the recent past. If the US Government wanted to get something big for trading Taiwan, during the Trump Administration was the last time to do it (during max tensions), and the trade should have been North Korea being dismantled (as is; meaning the Kim family goes, the nukes go; only China could make that happen) and reunited with South Korea. The US Govt no doubt thinks their bargaining position is stronger than it actually is, and so doesn't think they needed to make such a trade, instead the US will get nothing and Taiwan will end up in China's hands regardless and North Korea will remain a key China pawn.
I think China is happy to play the long game, since every passing year strengthens their position and weakens Taiwan's. But who knows what will happen with the next CCP succession.
Every passing year brings China one step closer to demographic collapse, the inevitable consequence of the One Child policy, and China is well aware of this.
Japan is several decades ahead of China on the same path, its population has been declining since 2011, and they have not imploded yet.
CCP requires economic growth for popular support, this is one reason for the Belt and Road Initiative and other dumping policies, gotta keep those factories running at all costs. If growth stagnates or declines and unemployment spikes, things will get bloody.
Also China's demographic dividend will end while it's at a much lower stage of development than Japan in the 80s/90s. And unlike Japan, China's managed to piss off pretty much every major power who might have been inclined to help them.
Very improbable. Nobody cares enough about Taiwan to even officially recognize its existence.
Taiwan was the United States' 13th largest goods export market in 2019.
https://ustr.gov/countries-regions/china/taiwan
If you think the US doesn’t care about a Top 10 trade partner, and in particular the trade partner that houses TSMC, well have I got news for you. Do we care enough to go into a full-on firefight immediately? I don’t know. But we care enough that it would be a massive diplomatic row, and we’d almost certainly have a naval peacocking as a show of force. And nobody wants that right by South Korea, India, and Japan.
Why do you think the US is working with TSMC to build plants in the US and intel is building more facilities as well? It's fairly obvious nobody has any confidence in the long run that Taiwan will remain independent.
I assure you we have not seen the last of war. Ignoring or minimizing the possibility is self-destructive. You must acknowledge the possibility of disaster before you can engage in the active work of averting it.
Never even said that. Just saying the Taiwan situation is not going to start a massive conflict.
Additionally, world wars are never triggered intentionally. World War I broke out essentially because the decision makers ignored the wider consequences of +1'ing their opponent's military preparations (one author characterizes it as Europe sleepwalking into war). World War II broke out unexpectedly because Hitler underestimated the spine the Allies had when he requested they add Danzig to his list of silver platters. World War III will likely break out when someone crosses a red line to one party that isn't clearly communicated to another party--and a Chinese attempt to invade Taiwan is one of the likelier options on that list.
The only reason Taiwan "insists" on that is because of the threat of invasion if they declare independence. Just because a bully tells you to stop hitting yourself while they smack your hand into your face does not mean the bully is not doing the hitting.
Before this, the bully was the Taiwan government. Then they lost the mainland, and fled to Taiwan, where they got sheltered by the Americans.
And now, you’re saying that the government on the mainland is now the bully over little Taiwan.
But the fact remains, that the last stage of the Chinese Civil War is not over.
History is useless since none of the previous world wars had nuclear powers at play. You can't compare before and after. Like, not at all.
China currently lacks sufficient amphibious assault ships to actually move enough troops to take control of the island. They certainly can bomb the crap out of it, but they won't have the resources necessary for boots on the ground for some time still.
They're building ships fast, but it's still something like a 15 year horizon before they'll have sufficient material, perhaps longer. At the least we can expect them to wait until the 2 carriers they're constructing are done.
Most credible analysts I've read believe a serious attempt at invading the island won't come until Xi's successor. Rick Joe is good to read on this topic. He writes for the major media outlets, as well as posting as PLARealtalk on reddit.
India is making strides but has internal issues and both China and Pakistan as border issues. South Korea has been under threat from the North for far too long.
It just makes financial sense to spread out to more stable regions of the world and to take the money offered when it shows up.
I figure we are still a decade or two out before Africa comes online and South and Central America will need to calm down to get good investment to spread about.
I am surprised Canada is not trying to get more of the action.
And I wil not be suspired if future POWER and Z are all Intel Fabbed.
2. Those Fabs are not leading edge and not target for the general consumer brands. The current target of 5nm in ~2024, a time when TSMC should be on 2nm GAAFET in their main Fab in Taiwan.
3. The only real reason is National Security, which is now being handled by Intel. These chip were used used and Fabbed by IBM and GF before GF's exit on leading edge.
4. 60% of TSMC's customers are US companies, but these US companies 's customer aren't 60% Americans. And not all Americans chips needs to be Fabbed locally.
I don't think the DoD is willing to put all of it's eggs in that one basket that hasn't been performing lately. Particularly when the DoD sees massive amounts of edge inference being a key weapons component going forward and they need newer nodes than 14nm+++++++.
Hard to blame them.
That is of course, as you said unless both Intel and Samsung seriously messed up their plan.
fucked, because we already have a foundry on US soil suckers. The whole point of TSMC being an ace card towards guaranteeing US defense coverage goes away the second you give it up. Ukraine is a perfect real life example https://en.wikipedia.org/wiki/Budapest_Memorandum_on_Securit...
It's certainly less risky for Taiwan for TSMC to be the one building and running the most high profile element of it than if it was someone else.
EDIT: not the best article but deals briefly with Arizona semiconductor history toward end of story article: https://www.allaboutcircuits.com/news/semiconductor-fabs-dot...
It's partially to mitigate supply risk. Recently, drought has worsened the supply of chip production, many of which are concentrated in Taiwan[1]. If you go on any PC parts forum, you will see that parts like new GPUs are nearly impossible to get. Some years back, flooding in Thailand[2] also threatened the supply of hard drives.
[1] https://thediplomat.com/2021/03/taiwan-prays-for-rain-as-dro...
[2] https://www.theguardian.com/technology/2011/oct/25/thailand-...
Tell these to people who keep mining at loss for god knows what reason
(Yes, if you are really betting on that, you may as well skip the mining and buy the coins outright, but maybe they have cheap power or other reasons)
> Industry statistics indicate that creating an integrated circuit on a 300 mm wafer requires approximately 2200 gallons of water in total, of which 1500 gallons is ultra-pure water.
Its honestly just overdue and this is as good as time as any to build this factory.
Taiwan was good enough. Now things are getting hairy and these companies have gotten massive and the demand for chips has gotten greater.
Time to build.
Even ignoring the political reasons, just consider what happened to WD and Seagate during the floods in Thailand in 2011.
We're talking about a a revisionist, expansionist state that is rising, believes the existing power to be fragile, weak, and on the decline. It isn't a question of "Viability". The CCP sees serious geopolitical benefits to an armed conflict.
That's a false dillema. The choice is not between self-sufficiency and war with China, and dependency and ever-lasting peace with China.
The choice is actually between war with China when China controls all the industry, and war with China when it fails to control and restrict access to that resource.
Given the choice, achieving self-sufficiency is more likely to ensure peace/dissuade war than dependency.
Why would a war with China be inevitable?
It's speculation, but China appears to be building up normalcy bias in Taiwanese defenses, get them used to seeing Chinese aircraft and making them more complacent.
Add the fact that China's national victim mentality is driving it to restore itself to it's perceived state before the "west" messed it up (includes regaining Taiwan), and that objective being one of the socio-political foundations of the CCP's power, and the CCP being well aware that China will start running out of young people in next decade due to one child policy (makes military adventures harder for both manpower and economic reasons), and Xi consolidating more power under himself and his cult of personality than any Chinese ruler since Mao...
Yeah if I was Taiwan I'd be subtly building some extra "tourism infrastructure" on the beaches and quietly stepping up training requirements for reserves; and give the US whatever it wants in exchange for a couple of carrier battle groups if needed.
It's funny because anybody with a brain could look at China in 1970 and look at it now, while asking themselves: "Hmmm, how did China do with that opening up to the US/West?"
It's laughable, but anti-US will say "buh buh buh... IRAQ!" as some sort of blanket example that the US has to stand aside now and let anybody do anything they want.
Reducing three countries to rubble (Iraq, Afghanistan, Libya) in just the past 20 years means exactly that the US has no place in international politics and should be treated as an extremely dangerous country, sanctioned by the United Nations for warcrimes, made to pay trillions in damages and slowly be re-introduced into the global village once they are demilitarized.
Iraq - I didn't agree with it. But look at the country now versus 20 years ago... it's a democracy in much better shape.
Afghanistan - we needed to take out al-Qaeda after the attacks. The country is broken, and outside actors help the Taliban continue their terrorism. You support the Taliban? Popular opinion supports the democratic government.
Libya - Russia and Turkey are the big players there now, continuing the civil war. I guess it's not as simple as you make it out to be, huh?
And yet you gloss over the countries like Japan, South Korea, Phillipines, Colombia, Eastern Europe and Western, that have all entered our sphere after war and have done amazingly economically while retaining high quality democracies.
Without the US, the world would quickly devolve back to fighting warlord/autocratic ethnostates. Or an autocratic CCP-controlled world.
Or... it's much more complicated and China has been slowly angling to usurp control (unjustly) for decades.
https://www.quora.com/What-you-think-about-Golden-Arches-the...
Besides, most of the US wars in the last half century have been started because the US was dependent on another nation. It seems a self-sufficient US may actually be less likely to go to war.
Definitely, but it's also a sensible precaution against natural disaster (earthquakes, tsunamis), madman-going-mad (especially Kim Jong Un, whose missiles have more than enough capability to reach Taiwan) and other (like the Suez Canal fuck-up) risk.
It simply is too much risk when too many companies depend on one single other company's (TSMC) products... Apple, AMD/ATI, almost all automobile supply chains...
Source? Because he's not by his actions even close to a communist.
And having been in China lately, I can ensure you CCP is only communist in name.
All of these companies are building in an area where it isn't clear if things will stay livable for the next 30 years.
Global integration of highly specialized industries not particularly dependent upon labor cost is not at all obvious--especially in the face of increasing political instability.
Part of the theory of engaging with China is that pulling it into the world would make them better worldwide citizens.
This has not worked. It has simply made China a bigger bully. So be it.
A gigantic amount of semiconductor manufacturing lives within 400 miles of China. That can't possibly be defended effectively against military action short of moving it a lot further away from China. That's what we're seeing the start of here.
Huge failure. They should've squeezed them in 1990-1992 as the USSR was dissolving and there was no pressing power to contend with. And North Korea they should've attacked then as well. No other way to solve this when the opponent will use your goodwill against you to play you for decades straight like the Kim dynasty.
Btw that gigantic ... Other than the Netherlands part, japan, South Korea and Taiwan are all keys here and close to a mad man and a sharp power.
Also noteworthy to mention all these are partially due to the failure of intel more than anything else. If intel keep arm and maintain the same level of chip production.
Ultimately we need at least two sources of critical components for any process rely upon them to be safe whatever the process is.
Props to economists on having zero foresight ability
Also, I'm not convinced that maximum efficiency is good for everyone. The US had a lot more manufacturing jobs before outsourcing to China. Sure, stuff is cheaper now, but a bunch of people no longer have decent paying jobs, and the labor market is now minimum wage or requires a college degree. I those people formerly doing manufacturing are a good deal poorer thanks to global integration. And we all are going to be poorer if all the economic-losers elect populists, since I can't think of any populists who didn't ruin their country if in power long enough.
Furthermore, I don't understand the fetishization of factory work (that both the left and right seem to have). It's menial, tough, body-breaking work, with long hours in sweaty, dangerous conditions. It's main appeal is to very poor nations where it is less awful than the alternative they had anyway (substance agriculture or starving). It's a good thing that we can become more productive and have higher paying safer jobs in the west and that in 'third world nations', they can find productive work that is safer and higher paying than their previous situation. The US had a lot of factory work when it was a poor nation, but there is no good reason to return. It's the concept of comparative advantage, and it makes us all better off on net.
A ship gets stuck in the Suez channel.
Let's drill a new channel ...
Which is actually what Egypt is doing, parts of the Canal are already redundant. The Ever Given had the bad luck of getting stuck in a non-redundant section.
Lots of talk about this last week:
https://news.ycombinator.com/item?id=26573236
Der letzte macht das Licht aus....
A few months ago there was a newspiece posted here in HN about a strategic billion-euro EU program that was aimed at subsidizing chip production capacity.
https://digital-strategy.ec.europa.eu/en/news/member-states-...
There is still the brain power, knowledge and money to compete with the US and China, the will and foresight is lacking. It is totally unclear that the EU could still pull this off in 30 years.
For example, ASML is an european company.
Jump-starting a manufacturing hub might be all it takes to be self-sufficient throughout the whole process.
AI, Crypto, VR, Robots ... not a single market leader is located in the EU.
I find it amazing that after missing the internet there is no discussion in Europe on how to avoid the same fate with artificial intelligence. There is only one European website in top 100 websites. And that is an Hungarian porn site.
AI will be 10x bigger than the internet. Yet it seems Europe does not even have it on the radar.
VR has only a few niche applications in entertainment. Crypto is a solution in search of a problem, and a gambling platform in the meantime.
AI isn't driving our cars, handling our customer service, or analyzing our business data. Good old fashioned algorithms written by humans are now being branded by marketing as "AI" which demonstrates we don't know what AI really is and we can't fend off its redefinition.
Robots is a tricky one because it's such a broad topic. Is an arm on an assembly line a robot? What about complex farming machinery? Does it have to look anthropomorphic?
From my bearish perspective on the first three industries, Europe and the EU (these are not interchangeable terms) have a pretty good track record this last decade compared to SV
Cypto = Switzerland basically the home of crypto if it has one
VR = Improbable. Multi $B from softbank
Robots = Europe is far ahead of the US in robotics, ABB / Universal robotics etc.
Universal Robots is owned by Teradyne, also in the US.
ABB I don't know. How are they market leaders?
Also, looking at ABB it seems this are those old industrial robots with few degrees of freedom. The future will be less specialized robots with more degrees of freedom. Powered by AI, not by explicitly coded motions.
Besides, it's fair game. The US had one of the two global leaders in AgTech in Monsanto, which was captured by the Germans. That repeatedly happens in biotech as a prominent example, most of the European biotech industry was built up on the back of acquiring US biotech companies.
Even ARM Holdings was originated heavily by US companies. There is hardly a piece of the global tech market that wasn't directly financed or made possible by the US tech industry and its mountain of profits and VC money, from SoftBank to China's tech companies to Shopify or Atlassian.
Aren't the EUV lithography machines that all advanced semiconductor manufacturers are dependent on made by a Dutch company?
If something is lacking in Europe it is:
* Scale
* The idea to dream big
Ehhhhh, the machines that are going to be in the chip factories mentioned here for example (ASML).
/s
And who says TSMC has peaked? Surely it’s best for everyone if Intel are kept under pressure.
Lıke ın a normal year TSMC would make lots and competıtors fabs would be more empty